Contents and Methods

Development of MRI technologies for quantitative measurements of oxygen consumption and vascular reactivity; applications to the study of brain energetics

Energy metabolism is both a prerequisite and a constraint for the development and maintenance of brain function. Even partial disruptions in the energy supply chain lead to immediate changes in brain function and, if prolonged for even a few minutes, to permanent damage or death.

We use the Davis model for BOLD calibration. The BOLD (Blood Oxygenation Level Dependent) signal is generated by variations in deoxyhaemoglobin concentration, which can be of vascular or neuronal origin. The neuronal component reflects the rate of oxygen consumption (CMRO2) and although it cannot be measured directly with non-invasive fMRI techniques, it can be isolated by calibrating the BOLD signal. The BOLD signal can be described in terms of CBF (cerebral blood flow), CBV (cerebral blood volume) and CMRO2 by introducing a calibration factor M. Since CBV can be approximated from CBF, and CBF and BOLD can be measured directly, the calibration factor M can be derived by performing two measurements, one of which introduces a manipulation that varies the relationship between CBF and BOLD. We will therefore develop a technique based on the administration of small doses of CO2 (5% in air), whose vasodilatory properties we will exploit. In addition to deriving CMRO2 measurements, we will also use CO2 administration to quantify CerebroVAscular Reactivity (CVR), an index of vascular compliance (ability of venous vessels to reversibly dilate). CVR will be determined simply as the relative change in BOLD and CBF signal during CO2 administration.

CMRO2 measurements will be associated with spectroscopy measurements to characterize the energetics of perception. We have recently shown that visual perception induces a decoupling between functional response and metabolic response; we aim to verify whether this decoupling can be associated with a different regulation of aerobic metabolism (CMRO2), which would have important consequences on the interpretation of functional data and on the understanding of pathologies or conditions that impact perception (for example, hallucinatory states).

This section is partially financed by the Lazio Region (NBP and FISASMEM projects)

Development of 23Na heteronuclear imaging for the investigation of early functional alterations in Alzheimer’s disease (AD)

AD is the most common type of dementia (80% of the total). It commonly occurs in the elderly, causing a progressive decline in cognitive domains, including attention, learning, memory and planning skills. AD has high and growing human and social costs. The aetiology of AD remains unknown. The diagnosis of dementia is mainly made on a clinical basis, in the absence of appropriate biomarkers that can provide an unequivocal diagnosis and characterise in vivo the metabolic and microstructural events associated with the early development of the disease.

Figura 1: Mappa del contenuto mielinico (ottenuta mediante tecnica T1/T2); si tratta di uno dei dati quantitativi MRI che saranno inseriti dello studio multiparametrico su AD

Figure 1: Map of myelin content (obtained by T1/T2 technique); this is one of the quantitative MRI data that will be included in the multiparametric study on AD

We will develop and exploit novel MRI techniques based on 23Na imaging in combination with quantitative MRI to identify potential disease biomarkers and explore the pathophysiological processes underlying microstructural tissue damage and cognitive impairment. Sodium plays a fundamental role in many physiological and biochemical functions. In particular, sodium homeostasis is associated with neuroinflammation, with potential sensitivity to vascular and metabolic alterations. Currently, no noninvasive MR imaging tools are available to detect neuroinflammation reliably. Therefore, we will develop heteronuclear MRI to study AD-associated neuroinflammation and as a component of a multiparametric quantitative MRI protocol to disentangle, in vivo and noninvasively, the neurophysiological alterations underlying neuroinflammation.

Figura 2: Prima immagine quantitativa di 23Na ottenuta come dato preliminare e relativa curva di calibrazione (ottenuta col metodo dei fantocci nel FOV, tre dei quali sono visibili attorno al cranio del volontario) 

Figure 2: First quantitative image of 23Na obtained as preliminary data and related calibration curve (obtained with the phantom method in the FOV, three of which are visible around the volunteer’s skull)

Characterization of the dynamics of brain networks and identification of components of non-neuronal origin.

The study of brain connectivity, based on the spatio-temporal characterization of the synchrony of BOLD signal fluctuations, is continuously expanding its fields of application, for example towards the early identification of neurological or psychiatric pathologies. Connectomic analysis is based on the characterization of differences compared to a reference, whether changes induced by a pathology, or simply the statistical comparison with a cognitively different condition. This is a complex procedure and subject to false positives. In fact, it should be remembered that connectomic analysis techniques, being based on the appreciation of the covariance structure of the data, are axes sensitive to coherent spurious signals, including the so-called “physiological noise” (i.e. the variations induced by physiological rhythms such as breathing , movement or heartbeat). The relationship between plastic modulation of networks and behaviour is a question of the utmost importance at the level of basic knowledge of brain function and the implications for the understanding of the main neurological and psychiatric pathologies. Our group is among the first to have addressed the topic of dynamic modulation of brain networks induced by brain function. In particular, we confirmed that the topology of brain networks at rest is globally preserved when executing a continuous cognitive task. We will continue to apply the techniques we have developed to identify the behavioural correlates of network dynamics, and in particular, we will extend our studies (performed using working memory as a model) to other cognitive domains (autobiographical memory, sensorimotor system).

Furthermore, by associating the CVR measurements developed in parallel (see above), we will continue in the development of denoising techniques aimed at separating the non-neuronal signal. This separation is important first of all to focus network studies on the true functional component; however (as a by-product), we believe that fluctuations of vascular origin can provide helpful information, particularly on sympathetic function. Over the three-year period we also aim to carry out a study on the variability of the vascular signal associated with ageing, an important confound in studies on the evolution of brain networks during ageing.

This section is partially financed by the Lazio Region (FISASMEM project)

​​Investigation of the role of glycogen in maintaining energy homeostasis

Our group has traditionally combined the experimental study of the energetics of the brain with its framing in computational models, which allow a more rigorous interpretation of the results by the integration of measures of different origins.

In collaboration with the Universities of Yale and Minnesota, we have formalised in detail the hypothesis that glycogen is essential for ensuring the availability of glucose in neurons. Glucose is a necessary substrate for some critical processes such as modulation of action potentials, axonal transport, filling of synaptic vesicles for neurotransmission. Glycogen is a reserve of glucose present only in the extensor muscles. Specifically, we introduced a hypothesis (GSG, Glucose sparing by Glycogenolysis) that astrocytic glycogenolysis plays a critical role in increasing the availability of glucose for neurons. Our modelling has shown that the GSG model is able to explain all the major experimental results on the subject, which are difficult to reconcile in the absence of GSG. The GSG model has the potential to provide a coupling mechanism between the electrical activity of neurons and the metabolic support of astrocytes. These processes are important for memory formation and consolidation and are altered during ageing. We plan to extend GSG modelling to include homeostatic mechanisms of pO2, pCO2, pH, with obvious synergies with CVR measurements and metabolism in vascular dementia.

Figura 3: Accordo tra misure sperimentali e previsioni dal modello GSG su pH, pCO2 e CMRO2 

Figure 3: Agreement between experimental measurements and predictions from the GSG model on pH, pCO2 and CMRO2